EP1215592A2 - Method for generating object identifiers, particularly for databases - Google Patents

Abstract

Method for generating object identifiers (OID) for processing, management and linking of object orientated data, in which the object is described using a single valued natural or surrogate key that has a corresponding object identifier. The natural object describing key (S1) is used to generate an artificial key (S2) using a hash-algorithm that is then used as the object identifier (S3). An Independent claim is made for a system for processing, management and joining together of object related data.

Description

The invention relates to a method for generating Object identifications for processing, administration and linking data associated with objects, being an object clearly through a possibly composed, natural and / or artificial key is described, the in turn, an object identification is assigned. Farther The invention relates to a system for processing, Management and linking of data assigned to objects with data processing means for generating object identifications and a corresponding computer program (product) to carry out the process.

The method or system mentioned comes during implementation a so-called data warehouse or operational Data store for use. This is where the problem arises, everyone Objects (rows in relational databases) with unique Keys. Such keys are needed to clearly refer to an object and to that object access assigned data as well as various data to be able to link (for example with database queries). For example, different data become one Object group depending on the data type in different databases / tables stored. To link to an object data belonging to various such tables must be in the table and the associated data for each table be identifiable. This can be natural or artificial keys (natural keys or surrogate keys) be used.

A natural key denotes an object using natural parameters that clearly identify the object. Are the objects physical or technical systems, so can be a natural key the various state variables of the system his. Objects can also be customers, sales products, Stock options and much more. At a Stock option could be the primary key designating the object composed of four natural components be: option type (call, put), base instrument (share), expiration month, Strike price. For a series of measurements for determination the hardness of workpieces could be a compound natural Key from the components: test type, Art of the workpiece, workpiece properties exist. The primary key is used to uniquely identify an object and to distinguish it from other objects.

As an alternative to natural keys, artificial ones are used Keys are often preferred. Every object will clearly assigned an ID. Be in a lookup table the assignments of the object identification (OID) to the for example by using composite natural keys objects described. In the different, dates to the databases / tables containing the objects then the keys describing the object (for compound Keys one column per component) replaced by the associated column of object identifications become. The OID thus becomes the primary key of the Master data file that identifies an object. Should a new object will be added to the database first the lookuptable loaded and checked that this object already exists. If not, it will generates an object identification (e.g. integer values, which are counted up) and a new entry in the lookup table. Only now can with the to proceed with the new object. However, it must be ensured be that the entry of the new object and its OID for the entire system (company) is clear, i.e. Double assignments must be excluded. To do this, a commit must occur after each insert, whereby the new values are finally written to disk, after which other processes to the changed values can access.

This procedure is very time-consuming and time-consuming expensive.

With the described method, you also have to query each database (Linking object data from different Tables (join)) the associated object identifications can be taken from the lookup table. This process increases the processing time considerably.

Instead of the composite natural key you can default IDs are also used (e.g. security key figure, Customer number, etc.). However, this occurs frequently practical problems because these IDs are often unique Labeling is not sufficient and then additional artificial numbers must be added. Are common such IDs also have historical reasons and need to be changed be subjected to current needs no longer match. To simplify, you resort therefore, here also back to artificially generated OIDs.

Either integer values are used as object identifications used by counting up or by generating Random numbers arise, or it becomes time stamps or Functions of this used. Basically, these require Methods the implementation of table lookup procedures, which is the relation between natural and artificial Reflect keys. This is a great overhead connected who is involved in the generation of the OIDs as well as in Database queries limit speed and increase costs effect.

The object of the present invention is therefore a method for generating object identifications in systems of the Specify the type described above, avoiding the Disadvantages mentioned above new entries and queries in databases made possible in a shorter time than before.

This object is achieved by the features of patent claim 1 solved. An inventive system for implementing the method is given by claim 5. The procedure leaves execute by a computer program according to claim 6. Advantageous configurations result from the respective Subclaims and the following description.

According to the invention, the description of the object Artificial key using a hash algorithm Generated key that used as object identification becomes. For example, a composite natural key generates a hash code and as an OID used. But it can also consist of a compound natural and artificial keys or from a pure artificial key using a hash algorithm new artificial key are generated, the inventive Object identification is used.

Hash algorithms are often used in the field of digital signatures used to convert documents of variable length into one To hash code that transforms the content of the document into compressed and non-retroactive form. The hash code is then the input value for the digital one Signature. The advantages of hash algorithms are big Throughput, d. H. the fast processing of large amounts of data, and collision resistance, d. H. to a given one The input string (document) can hardly be changed String (document) can be found that is the same Hash code. Examples of hash algorithms are MD5 (512-bit input blocks, 128 hash code length Bit), the standardized ISO / IEC 10 118-2 or SHA-1, as FIPS 180-1 standardized (512-bit input blocks, hash code length 160 bit). In practice, natural keys can be any (input) length can be encoded.

It is advantageous that the hash codes have fixed lengths, so that the object identifications according to the invention also have fixed lengths. The preferred length from which the preferred hash algorithm also depends on that be measured that at a required average or maximum number of generated object identifications the likelihood of collisions becomes so small that in practice it does not It can be expected that the same OIDs for different objects be calculated. Of course, for that The inventive method also hash algorithms be designed using a hash code of the desired length produce.

It has proven advantageous to use SHA-1 as the hash algorithm. It can be estimated that if 10 ¹⁰ OIDs are generated per year, an average collision occurs every 10 ¹⁴ years. This probability is so low that one can speak with a probability bordering on certainty of a unique assignment of the hash code used as an OID to an object.

This clear assignment of an object to its object according to the invention Object identification makes it possible for everyone Avoid lookup tables. According to the invention generates its OID for the object itself, the conventional own Generation of OIDs and subsequent assignment of the OIDs to their objects can thus be omitted. Even with database queries can respond to time-consuming queries from natural Keys or on a table lookup procedure for the determination the associated object identification can be dispensed with. This has proven to be enormously time-saving, so that the disadvantage of the usually increased space requirements (compared to using e.g. integer values as OIDs) hardly matters.

The generation of an object identification using SHA-1 can in a specific example within 20 microseconds take place what access times to a lookup table clearly falls short.

An inventive system for processing, management and linking data associated with objects contains data processing means to create a the object object identification associated with the descriptive key, According to the invention, this data processing means Computing unit, from which the object describing An artificial key using a hash algorithm Generate keys as object identification. On The computer program according to the invention contains program code means, to carry out the method according to the invention if the program on a computer, in particular the computer unit of the system according to the invention is processed. The program can be stored on various storage media such as EE-PROMs Flash memories, but also CD-ROMs, floppy disks or Hard drives.

The computer unit of the system according to the invention is calculated by means of the computer program mentioned (also a hardware solution would be conceivable) from an input data stream that which contains the key describing the object, the associated one Hash code that is output as object identification becomes.

Figur 1

zeigt ein Datenflussdiagramm zum Erzeugen und zum Laden von Objektidentifikationen (OID) nach herkömmlicher Art.

Figur 2

zeigt ein Datenflussdiagramm zum erfindungsgemäßen Generieren von Objektidentifikationen.

Figur 3

zeigt das Datenmodell für eine Messreihe mit Messwerten eines Werkstücks zur Schilderung einer Anwendung der Erfindung.

In the following, the invention and its advantages are to be explained in more detail with reference to the attached figures.

Figure 1

shows a data flow diagram for generating and loading object identifications (OID) in a conventional manner.

Figure 2

shows a data flow diagram for generating object identifications according to the invention.

Figure 3

shows the data model for a series of measurements with measured values of a workpiece to describe an application of the invention.

The steps (S1-S7) are outlined schematically in FIG. 1, that go through in a conventional process need to be loaded to load an object identification (OID) or generate new. It is said to be composed of a natural Key to be run out of a particular one Object (e.g. stock option or a workpiece) clearly identified and differentiated from other objects. Such a primary key could be with a stock option composed of four natural components be: option type (call, put), base instrument (share), expiration month, Strike price. With a workpiece that one If the series of measurements is subjected, the composite could be natural Key from the workpiece designation, the manufacturing date as well as other parameters, such as type designation, Manufacturer, name of the final inspector, etc., exist.

Is that supposed to be identifiable by its natural key? Object found by a database query or in the database must be included, the natural Key of the object set up (S1) and a table lookup procedure can be entered (S2). An existing one Lookup table is searched for the natural key and the associated OID is determined (S3). Is located a corresponding entry in the lookup table, so read the OID belonging to the natural key and used for subsequent database queries (S4).

If none belonging to the entered natural key OID can be determined from the lookup table a new OID will be assigned (S5). For example, the new OID in a simple way by counting up an integer to be generated. So that the newly created object identification (OID) in the entire system, for example one Company is available at the same time and Double assignments must be excluded after the insert (Insert a new object) commit (S6), whereby the new values are finally written to disk become. In the lookup table, the natural Key of the added object and its newly created one OID entered (S7).

As shown in Figure 2, the described shorten the complex process by the invention considerably. The natural key of an object becomes automatic the associated hash code is always generated (S1, S2). According to the invention, this hash code represents the object identification (OID) of the object (S3) used for database queries or describing a database / table with the Data associated with the object is used.

When using the SHA-1 hash algorithm with a natural one Keys with an input length of 512 bits can be in in this case an OID of the within 20 microseconds Length 160 bits (20 bytes) can be generated. This increases the Processing speed compared to the conventional one The method of Figure 1 significantly.

A simple application of the invention is intended on the basis of FIG in connection with a series of measurements with measured values of a Workpiece are explained. With 1 are the measured values a series of measurements 2 referred to a workpiece 3. The parameters for measurement series 2 and workpiece 3 are in the entities / tables MEASURING SERIES and WORKPIECE separately filed (normalization). Each series of measurements has an assigned one OID as primary key, the same applies to the Measured values and the workpieces. As foreign keys occur as Columns of the entity MEASURED VALUES the OIDs of the associated Measurement series (OIDMR) and the workpiece (OIDWS).

A possible structure of the tables is as follows (indication of the Columns):

a) MEASURED VALUES:

OID

date

time

VALUE (measured value)

Status

OIDMR (OID of the measurement series)

OIDWS (OID of the workpiece)

b) MEASURING SERIES:

OID

description

beginning

The End

Carried out by ...

Other parameters

c) WORKPIECE:

OID

description

Made on ...

Checked by ...

Other parameters

Select MR.Bezeichnung, MR.Beginn, WS.Bezeichnung, MW.Datum, MW.Zeitpunkt, MW.Wert

From MESSWERTE MW, MESSREIHE MR, WERKSTÜCK WS

Where MW.Zustand = xxx and
   MW.OIDMR = MR.OID and
   MW.OIDWS = WS.OID

A database query (query) should select all measurements that resulted in a certain status xxx. In addition to the measured value (from the table MEASURED VALUES), some other associated parameters are to be output, namely the date and time of the measured values, the designation of the workpiece and the start and designation of the measurement series. This query can be represented as follows:

Select MR designation, MR start, WS designation, MW date, MW time, MW value

From MEASURED VALUES MW, MEASURING SERIES MR, WORKPIECE WS

Where MW.Status = xxx and
MW.OIDMR = MR.OID and
MW.OIDWS = WS.OID

All keys are implemented using OIDs according to the invention. These OIDs are made from the appropriate natural ones Keys or a combination of natural Keys and OIDs (foreign keys) generated. any natural keys appear as attributes in the tables on.

KURSE (Kurse der Optionen)

TRANSAKT (Transaktionen, wie Käufe/Verkäufe)

POSITIONEN (offene Positionen in Depots)

Another example of the application of the invention are the stock options already mentioned above, which are available in a database, for example in the following three tables:

COURSES (courses of options)

TRANSACT (transactions, such as purchases / sales)

POSITIONS (open positions in depots)

The columns of these tables are when using OIDs as follows:

a) COURSES

OptionsOID

option type

underlying instrument

expiration month

base Price

Current exchange rate

DatumAktuellerKurs

The OptionsOID is the primary key of the table and identifies an option. The natural key is only attributes of the table.

b) TRANSACT

Transaction ID

date transaction

buying selling

OptionsOID

number

BezahlterPreis

Commission

Etc.

The OptionsOID acts as a foreign key. The TransactionId is the primary key of the table.

c) POSITIONS

deposit number

OptionsOID

number

success

Etc.

The OptionsOID acts as a foreign key. The deposit number is the primary key of the table.

select K.AktuellerKurs from KURSE K, POSITIONEN P

   where P.OptionsOID = K.OptionsOIDThe determination of the current prices for all positions in a portfolio can be formulated as the following query:

select K.Current course from COURSES K, POSITIONS P

where P.OptionsOID = K.OptionsOID

The OptionsOIDs are also from the four-component natural key (option type, basic instrument, expiry month, Base price), e.g. B. (C, Bayer, Dec. 00, 40.00), generated using the hash algorithm SHA-1 and 20 Bytes large.

The invention ensures direct and quick generation unique object identifications and at the same time resulting high access and processing speeds with database queries.

Claims (7)

Method for generating object identifications for processing, managing and linking data associated with objects, an object being uniquely described by a possibly composed, natural and / or artificial key, which in turn is assigned an object identification,
characterized in that an artificial key is generated from the key describing the object using a hash algorithm and is used as object identification. Method according to claim 1, characterized in that the artificial keys generated as object identifications have a fixed length. Method according to Claim 1 or 2, characterized in that the length of the artificial keys generated is dimensioned such that the probability of ambiguous assignments occurring between the generated artificial key and the key describing the object, depending on the average number of object identifications generated, is sufficient becomes low. Method according to one of Claims 1 to 3, characterized in that the SHA-1 method is used as the hash algorithm, object identifications with a length of 20 bytes being generated. System for processing, managing and linking data associated with objects, in particular in databases, an object being clearly described by a possibly composed, natural and / or artificial key, with data processing means for generating an object identification assigned to the key describing the object, characterized in that that the data processing means have a computing unit in order to generate an artificial key as object identification from the key describing the object by means of a hash algorithm. Computer program with program code means for a method according to one of claims 1 to 4, if the computer program from a dedicated computer, processed in particular the computing unit according to claim 5 becomes. Computer program product with program code means that are stored on a computer-readable medium in order to carry out a method according to one of claims 1 to 4, if the computer program on a computer or a corresponding computing unit, in particular a computing unit according to claim 5, is carried out.

Images